Enhancing SUV Accuracy of PET/MR for Clinical Trial Qualification

Standardized Uptake Values (SUVs), normalized activity concentration, measured using PET/MR have inaccuracies ≥ 20% which exceeds National Cancer Institute / American College of Radiology Imaging Network (NCI/ACRIN), Radiological Society of North America / Quantitative Imaging Biomarkers Alliance (RSNA/QIBA) specifications and disqualifies PET/MR from multicenter or cooperative group clinical trials. High inaccuracy is primarily due to poor attenuation correction (AC) owing to lack of computed tomography (CT) data. This study will develop acquisition and analyses methods to synthesize CT images from MR data that can be used to achieve SUVs that are within 5% of those obtained using PET/CT (reference standard), thus meeting accuracy requirements needed to qualify for multicenter trials. The overall goal of this research project is to validate clinically practical methods for producing MR-based attenuation correction information which is needed to produce quantitatively accurate PET images from a PET/MR scanner. Existing commercial PET/MR systems use methods that are inaccurate.